Machine Learning and IP Address Features

Data Skeptic

A frequent use case for applying machine learning is predicting some outcome based on Internet traffic. Although some effort is required, a well formated web server log can provide a wealth of raw data to be transformed into useful features for learning. Every company with an internet presence has a web server. With the exception of high security sites that make a point of not logging, these relatively compact text files are often managed through log rotation that moves the history to a persistent store (S3 being a popular choice) where they are (relatively) inexpensive to keep and may live for a very long time without threat of archiving for sp

Next Episodes



Data Skeptic

The Data Refuge Project @ Data Skeptic

📆 2017-03-03 01:00